The system in which a microscope and CCD camera with a PC 486 computer are used to capture a pattern image and the electrical output from the CCD camera is converted by a VIGAS card to a digital signal which is proportional to the brightness of the pattern. The electronic noise which is caused by the video chain should be suppressed.

Behind we obtain the the film quality characteristics, a series of image processing methods is adopt to find the particles on the film. The subjects for filtering leave out the noise and enhance the signal part of an image. Here we use a low pass mask and Laplacian mask of 3 3 matrix to perform the filtering process[6-8]. We use the bilevel thresholding method, so the original image is divided into two pixel classes; that is the below-threshold pixel and the above-threshold pixel. From the binary image we can obtain the area PA of a particle. To find the edge of a particle to determine the grain size we use a gradient enhancement method. The Laplacian edge extraction does not always provide closed lines in an image which is not binary encoded. But here the image which is operated by Laplacian mask convolution without problem can be calculated to obtain the length PL of the edge of a particle. The circularity Rc can be define as 4PA /. The ratio Re which is derived from the maximum axis divided by the minimum axis of the micro-particle can help us to understand the problems of the particle on the holographic plate.

Then we recover the original image gray level exclude the pixels in the micro-particle

to calculate the film quality characteristics: ATR, MTR, TPATR and RMSTR. A new concept, which is deriving from the surface roughness definitions, the film quality characteristics are defined as follows:

(1). average thickness variation(ATR):

ATR= 

where, t = , the mean thickness is Kt.

K : a gray level to thickness constant depends on the optical system

f : input image

N : total number of pixels, N=mn.

f(x,y) : the gray-level value of pixel(x,y)

ATR describes, more closely than any other, the frictinal quality of a film at the spinning manufacturing process.

(2). maximum thickness variation(MTR):

MTR= K [ Gmax-Gmin]

where Gmax is the maximum gray level value and Gmin is the minimum gray level value.

MTR is the peak to valley depth of the film. It's easy to calculate but undesirable to rate a complete surface of a film only on the maximum peak and valley. We often use this value to judge if there is any micro-particles on the film.

(3). ten point average thickness variation(TPATR):

TPATR=K[ _ ]

where

G(A);G(B);G(C);G(D) and G(E) is the maximum five gray levels.

G(A');G(B');G(C');G(D') and G(E') is the minimum five gray levels.

TPATR can give reliable information on the characteristic profile of the film.

(4). root mean square thickness variation(RMSTR):

RMSTR=K

RMSTR is rarely used because of the calculation work is very large.

With these parameters for qualitative measurement of the quality characteristics of film, we can make a adjustment on the parameters in film fabricating process(Fig.2).

nextindexback